9:00 am - Introduction to LASI
9:15 am - DataQuest Lab - Vik Paruchuri
We'll be analyzing student data, and discussing design and data collection strategies for dataquest.io (http://dataquest.io/). First, we'll walk through an interactive lesson on the dataquest platform that analyzes anonymized student data and finds points at which students drop out, or take unexpected actions. We'll pause as we go through, and discuss why this might be happening, and try to understand the student mindset. After we finish the lesson, we'll look at the platform more broadly, and chat about what data should be collected, and what design choices might boost learner engagement.
10:15 am - Paper Prototyping Visual Learning Analytics Workshop - Kim Ducharme (http://www.cast.org/about/staff/kim-ducharme.html#.VXWuOGRViko), Ge Vue, Garron Hillaire (http://www.cast.org/about/staff/garron-hillaire.html#.VXWuU2RViko)
We will facilitate a process around shaping and refining both an idea for data visualization and a theory of change on how the design is attempting to setup administrators, teachers, and/or student interactions that result in a change of behavior. We will introduce what we mean by “theory of change informing design” and then move into a short introduction to paper prototyping by working in small groups on the actual development of a data visualization paper prototype. The workshop will conclude with a short discussion to consolidate the opportunities and problems in research on visualization for learning analytics.
DATA Resources (https://docs.google.com/spreadsheets/d/1j_7vyTv2fGQTZ9GlyYh-iDVCxvbMlcpJ-7OpI__25Mk/edit#gid=0)
12:15 pm - Break for Lunch
1:15 pm - Panel on Learning Sciences and Learning Analytics
Jie Chao, Chad Dorsey, Hee-Sun Lee and Charles Xie, The Concord Consortium
Learning analytics can provide important insights into learning in many domain-specific situations. When informed by research on learning, computational models and approaches can permit us to gain a richer understanding of how learners acquire and progress with scientific and engineering practices and content understanding. Panelists from the Concord Consortium will describe a variety of analytics approaches including process analytics applied to the engineering design process, text analysis of student scientific argumentation and a novel application of Bayesian Knowledge Tracing to student learning of a mechanical system in a game-like application. The panel will pay special attention to how information from these and related techniques can provide important feedback to aid teaching and learning and can contribute back to aid ongoing learning research.
2:00 pm - Panel discussion on MOOCs & Learning Analytics - Justin Reich (http://edtechteacher.org/team/justin/) (chair), Michael Yeomans, John Hansen, Henry Eyring (http://harvardx.harvard.edu/research)
One of the three goals of edX and HarvardX is to advance the science of learning. Both the data and platforms of massive open online courses afford exciting opportunities to examine student learning behaviors in real-time, to experiment with diverse approaches to instructional design, and to characterize the scaffolds that lead to student success and the barriers that thwart learning. In this panel, graduate students and post-doctoral researchers from four departments at Harvard will introduce some of their research projects, as they relate to time-series analyses, text analysis, performance dashboards, and census-based geo-location.